roo 2.3.0 → 2.10.1

data/lib/roo/excelx/cell/number.rb

@@ -1,37 +1,37 @@
+# frozen_string_literal: true
+
 module Roo
   class Excelx
     class Cell
       class Number < Cell::Base
-        ERROR_VALUES = %w(#N/A #REF! #NAME? #DIV/0! #NULL! #VALUE! #NUM!)
+        attr_reader :value, :formula, :format, :cell_value, :coordinate
 
-        attr_reader :value, :formula, :format, :cell_value, :link, :coordinate
+        # FIXME: change default_type to number. This will break brittle tests.
+        attr_reader_with_default default_type: :float
 
         def initialize(value, formula, excelx_type, style, link, coordinate)
           super
-          # FIXME: change @type to number. This will break brittle tests.
           # FIXME: Excelx_type is an array, but the first value isn't used.
-          @type = :float
           @format = excelx_type.last
-          @value = link? ? Roo::Link.new(link, value) : create_numeric(value)
+          @value = link ? Roo::Link.new(link, value) : create_numeric(value)
         end
 
         def create_numeric(number)
-          return number if ERROR_VALUES.include?(number)
-
+          return number if Excelx::ERROR_VALUES.include?(number)
           case @format
           when /%/
             Float(number)
           when /\.0/
             Float(number)
           else
-            number.include?('.') ? Float(number) : Integer(number)
+            (number.include?('.') || (/\A[-+]?\d+E[-+]?\d+\z/i =~ number)) ? Float(number) : Integer(number, 10)
           end
         end
 
         def formatted_value
-          return @cell_value if ERROR_VALUES.include?(@cell_value)
+          return @cell_value if Excelx::ERROR_VALUES.include?(@cell_value)
 
-          formatter = formats[@format]
+          formatter = generate_formatter(@format)
           if formatter.is_a? Proc
             formatter.call(@cell_value)
           else
@@ -39,46 +39,59 @@ module Roo
           end
         end
 
-        def formats
+        def generate_formatter(format)
           # FIXME: numbers can be other colors besides red:
           # [BLACK], [BLUE], [CYAN], [GREEN], [MAGENTA], [RED], [WHITE], [YELLOW], [COLOR n]
-          {
-            'General' => '%.0f',
-            '0' => '%.0f',
-            '0.00' => '%.2f',
-            '#,##0' => proc do |number|
-              Kernel.format('%.0f', number).reverse.gsub(/(\d{3})(?=\d)/, '\\1,').reverse
-            end,
-            '#,##0.00' => proc do |number|
-              Kernel.format('%.2f', number).reverse.gsub(/(\d{3})(?=\d)/, '\\1,').reverse
-            end,
-            '0%' =>  proc do |number|
-              Kernel.format('%d%', number.to_f * 100)
-            end,
-            '0.00%' => proc do |number|
-              Kernel.format('%.2f%', number.to_f * 100)
-            end,
-            '0.00E+00' => '%.2E',
-            '#,##0 ;(#,##0)' => proc do |number|
-              formatter = number.to_i > 0 ? '%.0f' : '(%.0f)'
-              Kernel.format(formatter, number.to_f.abs).reverse.gsub(/(\d{3})(?=\d)/, '\\1,').reverse
-            end,
-            '#,##0 ;[Red](#,##0)' => proc do |number|
-              formatter = number.to_i > 0 ? '%.0f' : '[Red](%.0f)'
-              Kernel.format(formatter, number.to_f.abs).reverse.gsub(/(\d{3})(?=\d)/, '\\1,').reverse
-            end,
-            '#,##0.00;(#,##0.00)' => proc do |number|
-              formatter = number.to_i > 0 ? '%.2f' : '(%.2f)'
-              Kernel.format(formatter, number.to_f.abs).reverse.gsub(/(\d{3})(?=\d)/, '\\1,').reverse
-            end,
-            '#,##0.00;[Red](#,##0.00)' => proc do |number|
-              formatter = number.to_i > 0 ? '%.2f' : '[Red](%.2f)'
-              Kernel.format(formatter, number.to_f.abs).reverse.gsub(/(\d{3})(?=\d)/, '\\1,').reverse
-            end,
+          case format
+          when /^General$/i then '%.0f'
+          when '0' then '%.0f'
+          when /^(0+)$/ then "%0#{$1.size}d"
+          when /^0\.(0+)$/ then "%.#{$1.size}f"
+          when '#,##0' then number_format('%.0f')
+          when /^#,##0.(0+)$/ then number_format("%.#{$1.size}f")
+          when '0%'
+            proc do |number|
+              Kernel.format('%.0f%%', number.to_f * 100)
+            end
+          when '0.00%'
+            proc do |number|
+              Kernel.format('%.2f%%', number.to_f * 100)
+            end
+          when '0.00E+00' then '%.2E'
+          when '#,##0 ;(#,##0)' then number_format('%.0f', '(%.0f)')
+          when '#,##0 ;[Red](#,##0)' then number_format('%.0f', '[Red](%.0f)')
+          when '#,##0.00;(#,##0.00)' then number_format('%.2f', '(%.2f)')
+          when '#,##0.00;[Red](#,##0.00)' then number_format('%.2f', '[Red](%.2f)')
             # FIXME: not quite sure what the format should look like in this case.
-            '##0.0E+0' => '%.1E',
-            '@' => proc { |number| number }
-          }
+          when '##0.0E+0' then '%.1E'
+          when "_-* #,##0.00\\ _€_-;\\-* #,##0.00\\ _€_-;_-* \"-\"??\\ _€_-;_-@_-" then number_format('%.2f', '-%.2f')
+          when '@' then proc { |number| number }
+          when /^(?:_\()?"([^"]*)"(?:\* )?([^_]+)/
+            proc do |number|
+              formatted_number = generate_formatter($2).call(number)
+              "#{$1}#{formatted_number}"
+            end
+          when /^_[- \(]\[\$([^-]*)[^#@]+([^_]+)/
+            proc do |number|
+              formatted_number = generate_formatter($2).call(number)
+              "#{$1}#{formatted_number}"
+            end
+          else
+            raise "Unknown format: #{format.inspect}"
+          end
+        end
+
+        private
+
+        def number_format(formatter, negative_formatter = nil)
+          proc do |number|
+            if negative_formatter
+              formatter = number.to_i > 0 ? formatter : negative_formatter
+              number = number.to_f.abs
+            end
+
+            Kernel.format(formatter, number).reverse.gsub(/(\d{3})(?=\d)/, '\\1,').reverse
+          end
         end
       end
     end

data/lib/roo/excelx/cell/string.rb

@@ -2,12 +2,12 @@ module Roo
   class Excelx
     class Cell
       class String < Cell::Base
-        attr_reader :value, :formula, :format, :cell_type, :cell_value, :link, :coordinate
+        attr_reader :value, :formula, :format, :cell_value, :coordinate
+
+        attr_reader_with_default default_type: :string, cell_type: :string
 
         def initialize(value, formula, style, link, coordinate)
           super(value, formula, nil, style, link, coordinate)
-          @type = @cell_type = :string
-          @value = link? ? Roo::Link.new(link, value) : value
         end
 
         def empty?

data/lib/roo/excelx/cell/time.rb

@@ -4,15 +4,16 @@ module Roo
   class Excelx
     class Cell
       class Time < Roo::Excelx::Cell::DateTime
-        attr_reader :value, :formula, :format, :cell_value, :link, :coordinate
+        attr_reader :value, :formula, :format, :cell_value, :coordinate
+
+        attr_reader_with_default default_type: :time
 
         def initialize(value, formula, excelx_type, style, link, base_date, coordinate)
           # NOTE: Pass all arguments to DateTime super class.
           super
-          @type = :time
           @format = excelx_type.last
           @datetime = create_datetime(base_date, value)
-          @value = link? ? Roo::Link.new(link, value) : (value.to_f * 86_400).to_i
+          @value = link ? Roo::Link.new(link, value) : (value.to_f * 86_400).round.to_i
         end
 
         def formatted_value
@@ -24,19 +25,19 @@ module Roo
 
         private
 
-        def create_datetime(base_date, value)
-          date = base_date + value.to_f.round(6)
-          datetime_string = date.strftime('%Y-%m-%d %H:%M:%S.%N')
-          t = round_datetime(datetime_string)
-
-          ::DateTime.civil(t.year, t.month, t.day, t.hour, t.min, t.sec)
-        end
-
-        def round_datetime(datetime_string)
-          /(?<yyyy>\d+)-(?<mm>\d+)-(?<dd>\d+) (?<hh>\d+):(?<mi>\d+):(?<ss>\d+.\d+)/ =~ datetime_string
-
-          ::Time.new(yyyy.to_i, mm.to_i, dd.to_i, hh.to_i, mi.to_i, ss.to_r).round(0)
-        end
+        # def create_datetime(base_date, value)
+        #   date = base_date + value.to_f.round(6)
+        #   datetime_string = date.strftime('%Y-%m-%d %H:%M:%S.%N')
+        #   t = round_datetime(datetime_string)
+        #
+        #   ::DateTime.civil(t.year, t.month, t.day, t.hour, t.min, t.sec)
+        # end
+
+        # def round_datetime(datetime_string)
+        #   /(?<yyyy>\d+)-(?<mm>\d+)-(?<dd>\d+) (?<hh>\d+):(?<mi>\d+):(?<ss>\d+.\d+)/ =~ datetime_string
+        #
+        #   ::Time.new(yyyy.to_i, mm.to_i, dd.to_i, hh.to_i, mi.to_i, ss.to_r).round(0)
+        # end
       end
     end
   end

data/lib/roo/excelx/cell.rb

@@ -11,7 +11,7 @@ require 'roo/excelx/cell/time'
 module Roo
   class Excelx
     class Cell
-      attr_reader :type, :formula, :value, :excelx_type, :excelx_value, :style, :hyperlink, :coordinate
+      attr_reader :formula, :value, :excelx_type, :excelx_value, :style, :hyperlink, :coordinate
       attr_writer :value
 
       # DEPRECATED: Please use Cell.create_cell instead.
@@ -40,19 +40,23 @@ module Roo
       end
 
       def self.create_cell(type, *values)
+        cell_class(type)&.new(*values)
+      end
+
+      def self.cell_class(type)
         case type
         when :string
-          Cell::String.new(*values)
+          Cell::String
         when :boolean
-          Cell::Boolean.new(*values)
+          Cell::Boolean
         when :number
-          Cell::Number.new(*values)
+          Cell::Number
         when :date
-          Cell::Date.new(*values)
+          Cell::Date
         when :datetime
-          Cell::DateTime.new(*values)
+          Cell::DateTime
         when :time
-          Cell::Time.new(*values)
+          Cell::Time
         end
       end
 

data/lib/roo/excelx/comments.rb

@@ -12,10 +12,10 @@ module Roo
       def extract_comments
         return {} unless doc_exists?
 
-        Hash[doc.xpath('//comments/commentList/comment').map do |comment|
+        doc.xpath('//comments/commentList/comment').each_with_object({}) do |comment, hash|
           value = (comment.at_xpath('./text/r/t') || comment.at_xpath('./text/t')).text
-          [::Roo::Utils.ref_to_key(comment.attributes['ref'].to_s), value]
-        end]
+          hash[::Roo::Utils.ref_to_key(comment['ref'].to_s)] = value
+        end
       end
     end
   end

data/lib/roo/excelx/coordinate.rb

@@ -1,11 +1,18 @@
 module Roo
   class Excelx
-    class Coordinate
-      attr_accessor :row, :column
+    class Coordinate < ::Array
 
       def initialize(row, column)
-        @row = row
-        @column = column
+        super() << row << column
+        freeze
+      end
+
+      def row
+        self[0]
+      end
+
+      def column
+        self[1]
       end
     end
   end

data/lib/roo/excelx/extractor.rb

@@ -1,15 +1,32 @@
+# frozen_string_literal: true
+
+require "roo/helpers/weak_instance_cache"
+
 module Roo
   class Excelx
     class Extractor
-      def initialize(path)
+      include Roo::Helpers::WeakInstanceCache
+
+      COMMON_STRINGS = {
+        t: "t",
+        r: "r",
+        s: "s",
+        ref: "ref",
+        html_tag_open: "<html>",
+        html_tag_closed: "</html>"
+      }
+
+      def initialize(path, options = {})
         @path = path
+        @options = options
       end
 
       private
 
       def doc
-        @doc ||=
-        if doc_exists?
+        instance_cache(:@doc) do
+          raise FileNotFound, "#{@path} file not found" unless doc_exists?
+
           ::Roo::Utils.load_xml(@path).remove_namespaces!
         end
       end

data/lib/roo/excelx/format.rb

@@ -1,49 +1,57 @@
+# frozen_string_literal: true
+
 module Roo
   class Excelx
     module Format
+      extend self
       EXCEPTIONAL_FORMATS = {
         'h:mm am/pm' => :date,
         'h:mm:ss am/pm' => :date
       }
 
       STANDARD_FORMATS = {
-        0 => 'General'.freeze,
-        1 => '0'.freeze,
-        2 => '0.00'.freeze,
-        3 => '#,##0'.freeze,
-        4 => '#,##0.00'.freeze,
-        9 => '0%'.freeze,
-        10 => '0.00%'.freeze,
-        11 => '0.00E+00'.freeze,
-        12 => '# ?/?'.freeze,
-        13 => '# ??/??'.freeze,
-        14 => 'mm-dd-yy'.freeze,
-        15 => 'd-mmm-yy'.freeze,
-        16 => 'd-mmm'.freeze,
-        17 => 'mmm-yy'.freeze,
-        18 => 'h:mm AM/PM'.freeze,
-        19 => 'h:mm:ss AM/PM'.freeze,
-        20 => 'h:mm'.freeze,
-        21 => 'h:mm:ss'.freeze,
-        22 => 'm/d/yy h:mm'.freeze,
-        37 => '#,##0 ;(#,##0)'.freeze,
-        38 => '#,##0 ;[Red](#,##0)'.freeze,
-        39 => '#,##0.00;(#,##0.00)'.freeze,
-        40 => '#,##0.00;[Red](#,##0.00)'.freeze,
-        45 => 'mm:ss'.freeze,
-        46 => '[h]:mm:ss'.freeze,
-        47 => 'mmss.0'.freeze,
-        48 => '##0.0E+0'.freeze,
-        49 => '@'.freeze
+        0 => 'General',
+        1 => '0',
+        2 => '0.00',
+        3 => '#,##0',
+        4 => '#,##0.00',
+        9 => '0%',
+        10 => '0.00%',
+        11 => '0.00E+00',
+        12 => '# ?/?',
+        13 => '# ??/??',
+        14 => 'mm-dd-yy',
+        15 => 'd-mmm-yy',
+        16 => 'd-mmm',
+        17 => 'mmm-yy',
+        18 => 'h:mm AM/PM',
+        19 => 'h:mm:ss AM/PM',
+        20 => 'h:mm',
+        21 => 'h:mm:ss',
+        22 => 'm/d/yy h:mm',
+        37 => '#,##0 ;(#,##0)',
+        38 => '#,##0 ;[Red](#,##0)',
+        39 => '#,##0.00;(#,##0.00)',
+        40 => '#,##0.00;[Red](#,##0.00)',
+        45 => 'mm:ss',
+        46 => '[h]:mm:ss',
+        47 => 'mmss.0',
+        48 => '##0.0E+0',
+        49 => '@'
       }
 
       def to_type(format)
+        @to_type ||= {}
+        @to_type[format] ||= _to_type(format)
+      end
+
+      def _to_type(format)
         format = format.to_s.downcase
         if (type = EXCEPTIONAL_FORMATS[format])
           type
         elsif format.include?('#')
           :float
-        elsif !format.match(/d+(?![\]])/).nil? || format.include?('y')
+        elsif format.include?('y') || !format.match(/d+(?![\]])/).nil?
           if format.include?('h') || format.include?('s')
             :datetime
           else
@@ -58,7 +66,6 @@ module Roo
         end
       end
 
-      module_function :to_type
     end
-  end 
+  end
 end

data/lib/roo/excelx/images.rb

@@ -0,0 +1,26 @@
+require 'roo/excelx/extractor'
+
+module Roo
+  class Excelx
+    class Images < Excelx::Extractor
+
+      # Returns: Hash { id1: extracted_file_name1 },
+      # Example: { "rId1"=>"roo_media_image1.png",
+      #            "rId2"=>"roo_media_image2.png",
+      #            "rId3"=>"roo_media_image3.png" }
+      def list
+        @images ||= extract_images_names
+      end
+
+      private
+
+      def extract_images_names
+        return {} unless doc_exists?
+
+        doc.xpath('/Relationships/Relationship').each_with_object({}) do |rel, hash|
+          hash[rel['Id']] = "roo" + rel['Target'].gsub(/\.\.\/|\//, '_')
+        end
+      end
+    end
+  end
+end

data/lib/roo/excelx/relationships.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'roo/excelx/extractor'
 
 module Roo
@@ -11,14 +13,20 @@ module Roo
         @relationships ||= extract_relationships
       end
 
+      def include_type?(type)
+        to_a.any? do |_, rel|
+          rel["Type"]&.include? type
+        end
+      end
+
       private
 
       def extract_relationships
-        return [] unless doc_exists?
+        return {} unless doc_exists?
 
-        Hash[doc.xpath('/Relationships/Relationship').map do |rel|
-          [rel.attribute('Id').text, rel]
-        end]
+        doc.xpath('/Relationships/Relationship').each_with_object({}) do |rel, hash|
+          hash[rel['Id']] = rel
+        end
       end
     end
   end

data/lib/roo/excelx/shared.rb

@@ -4,12 +4,15 @@ module Roo
     #          reduce memory usage and reduce the number of objects being passed
     #          to various inititializers.
     class Shared
-      attr_accessor :comments_files, :sheet_files, :rels_files
-      def initialize(dir)
+      attr_accessor :comments_files, :sheet_files, :rels_files, :image_rels, :image_files
+      def initialize(dir, options = {})
         @dir = dir
         @comments_files = []
         @sheet_files = []
         @rels_files = []
+        @options = options
+        @image_rels = []
+        @image_files = []
       end
 
       def styles
@@ -17,7 +20,7 @@ module Roo
       end
 
       def shared_strings
-        @shared_strings ||= SharedStrings.new(File.join(@dir, 'roo_sharedStrings.xml'))
+        @shared_strings ||= SharedStrings.new(File.join(@dir, 'roo_sharedStrings.xml'), @options)
       end
 
       def workbook
@@ -27,6 +30,10 @@ module Roo
       def base_date
         workbook.base_date
       end
+
+      def base_timestamp
+        workbook.base_timestamp
+      end
     end
   end
 end

data/lib/roo/excelx/shared_strings.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require 'roo/excelx/extractor'
 
 module Roo
@@ -11,24 +13,34 @@ module Roo
         @array ||= extract_shared_strings
       end
 
+      def to_html
+        @html ||= extract_html
+      end
+
+      # Use to_html or to_a for html returns
+      # See what is happening with commit???
+      def use_html?(index)
+        return false if @options[:disable_html_wrapper]
+        to_html[index][/<([biu]|sup|sub)>/]
+      end
+
       private
 
       def fix_invalid_shared_strings(doc)
         invalid = { '_x000D_'  => "\n" }
         xml = doc.to_s
+        return doc unless xml[/#{invalid.keys.join('|')}/]
 
-        if xml[/#{invalid.keys.join('|')}/]
-          @doc = ::Nokogiri::XML(xml.gsub(/#{invalid.keys.join('|')}/, invalid))
-        end
+        ::Nokogiri::XML(xml.gsub(/#{invalid.keys.join('|')}/, invalid))
       end
 
       def extract_shared_strings
         return [] unless doc_exists?
 
-        fix_invalid_shared_strings(doc)
+        document = fix_invalid_shared_strings(doc)
         # read the shared strings xml document
-        doc.xpath('/sst/si').map do |si|
-          shared_string = ''
+        document.xpath('/sst/si').map do |si|
+          shared_string = +""
           si.children.each do |elem|
             case elem.name
             when 'r'
@@ -42,6 +54,98 @@ module Roo
           shared_string
         end
       end
-    end
-  end
-end
+
+      def extract_html
+        return [] unless doc_exists?
+        fix_invalid_shared_strings(doc)
+        # read the shared strings xml document
+        doc.xpath('/sst/si').map do |si|
+          html_string = '<html>'.dup
+          si.children.each do |elem|
+            case elem.name
+            when 'r'
+              html_string << extract_html_r(elem)
+            when 't'
+              html_string << elem.content
+            end # case elem.name
+          end # si.children.each do |elem|
+          html_string << '</html>'
+        end # doc.xpath('/sst/si').map do |si|
+      end # def extract_html
+
+      # The goal of this function is to take the following XML code snippet and create a html tag
+      # r_elem ::: XML Element that is in sharedStrings.xml of excel_book.xlsx
+      # {code:xml}
+      # <r>
+      #   <rPr>
+      #      <i/>
+      #      <b/>
+      #      <u/>
+      #      <vertAlign val="subscript"/>
+      #      <vertAlign val="superscript"/>
+      #   </rPr>
+      #   <t>TEXT</t>
+      # </r>
+      # {code}
+      #
+      # Expected Output ::: "<html><sub|sup><b><i><u>TEXT</u></i></b></sub|/sup></html>"
+      def extract_html_r(r_elem)
+        str = +""
+        xml_elems = {
+          sub: false,
+          sup: false,
+          b:   false,
+          i:   false,
+          u:   false
+        }
+        r_elem.children.each do |elem|
+          case elem.name
+          when 'rPr'
+            elem.children.each do |rPr_elem|
+              case rPr_elem.name
+              when 'b'
+                # set formatting for Bold to true
+                xml_elems[:b] = true
+              when 'i'
+                # set formatting for Italics to true
+                xml_elems[:i] = true
+              when 'u'
+                # set formatting for Underline to true
+                xml_elems[:u] = true
+              when 'vertAlign'
+                # See if the Vertical Alignment is subscript or superscript
+                case rPr_elem.xpath('@val').first.value
+                when 'subscript'
+                  # set formatting for Subscript to true and Superscript to false ... Can't have both
+                  xml_elems[:sub] = true
+                  xml_elems[:sup] = false
+                when 'superscript'
+                  # set formatting for Superscript to true and Subscript to false ... Can't have both
+                  xml_elems[:sup] = true
+                  xml_elems[:sub] = false
+                end
+              end
+            end
+          when 't'
+            str << create_html(elem.content, xml_elems)
+          end
+        end
+        str
+      end # extract_html_r
+
+      # This will return an html string
+      def create_html(text, formatting)
+        tmp_str = +""
+        formatting.each do |elem, val|
+          tmp_str << "<#{elem}>" if val
+        end
+        tmp_str << text
+
+        formatting.reverse_each do |elem, val|
+          tmp_str << "</#{elem}>" if val
+        end
+        tmp_str
+      end
+    end # class SharedStrings < Excelx::Extractor
+  end # class Excelx
+end # module Roo